2) Measure a task’s cost
Multiple cost metrics
- time: how long does it take to perform the task?
- ideally, measure minimum time that user must spend
- no “think time”
- experienced user should achieve this minimum
- subtleties in handling periods where user waits for sys.
- impact: how does the task affect system availability?
- use availability benchmarks, distilled into numbers
- learning curve: how hard is it to reach min. time?
- this one’s a challenge since it’s user-dependent
- measure via user studies
- how many errors do users make while learning tasks?
- how long does it take for users to reach min. time?
- does frequency of user errors decrease with time?